A New Feature Information Enhancement Method Based on Matrix Transformation and Cubic Spline Interpolation for Train Positioning

The output accuracy of train video odometry depends on high-quality images. When preprocessing the original image, the common methods tend to focus on the estimation of the interpolation data, and the feature information of the source image cannot be well preserved. Aiming at the loss of feature information after source image preprocessing, an image feature information enhancement method based on matrix transformation and Cubic Spline Interpolation is proposed. The method firstly transforms the original image matrix through singular value decomposition, and uses the cubic spline method to interpolate the transformed image matrix, and reconstructs a high-resolution enhanced image from the interpolated matrix, The feature information of the source image is preserved in the process of image preprocessing. The simulation results show that compared with other standard interpolation methods, the enhanced image obtained by the proposed algorithm can better retain the edge structure and texture information, and obtain better image enhancement effect. Based on the quantification criteria for objective evaluation of image quality, the enhanced image can improve the signal-to-noise ratio and peak signal-to-noise ratio, and at the same time, the root mean square error is also lower than other standard interpolation algorithms.